Recent randomized trial data from the Women's Health Study (WHS) demonstrated no overall effect of low dose aspirin on the primary endpoint of major cardiovascular disease (CVD), including MI, stroke, and CVD death. This contrasts with a meta-analysis of previous trials, mostly in men, that showed a 28% reduction in coronary heart disease. Whether the discrepancies are due to gender or dose remains unanswered. While additional trial data would be necessary to fully address this issue, WHS data can be used to determine whether varying usage within the study is a determinant of future CVD events. We propose to use state-of- the art statistical methods to estimate the causal effects of actual aspirin use during the trial to determine if actual use is more predictive of later CVD, and whether higher actual doses used have a greater impact on CVD. In addition, we will use the updated time-varying covariables to examine whether they modify the effect of actual aspirin use during the trial using similar methods. While data for the vitamin E intervention in the WHS are more consistent with previous null trial data, we found a reduction in the secondary endpoint of CVD mortality, as well as effect modification by age. There may be confounding by intermediate events in these data, however, which could distort estimates from both randomized as well as observational analyses. We will thus apply the same causal methods to determine whether there is an effect of actual vitamin E use or dose on the development of CVD after adjusting for time-varying covariables, as well as exploring the presence of modification of the effect of vitamin E. A 2002 recommendation by the Preventive Services Task Force for primary prevention strongly recommended that clinicians consider aspirin chemoprevention for all adults at increased risk of coronary heart disease, including postmenopausal women. The true effect of aspirin in women, and of varying effects by dose, thus has major public health relevance. For vitamin E, data from the WHS support current guidelines stating that use is not justified for CVD reduction. The findings of reduced risk of CVD mortality, however, as well as effects by age, should be explored further to clarify the public health message. [unreadable] [unreadable] [unreadable]